Taking Temporality Seriously: Modeling History and the Use of Narratives as Evidence

Social scientists interested in explaining historical processes can, indeed should, refuse the choicebetween modeling causal relationships and studying history. Identifying temporality as the definingcharacteristic of processes that can be meaningfully distinguished as “history, ” I show that mod-eling such phenomena engenders particular difficulties but is both possible and fruitful. Narratives, as a way of presenting empirical information, have distinctive strengths that make them especially suited for historical scholarship, and structuring the narratives based on the model allows us to treat them as data on which to test the model. At the same time, this use of narratives raises methodological problems not identified in recent debates. I specify these problems, analyze their implications, and suggest ways of solving or minimizing them. There is no inherent incompatibility between—but much potential gain from—modeling history and using historical narratives as data. Across the empirical subfields of political sci-ence, we find in recent years a renewed andgrowing interest in “historical macro-analysis” (Katznelson 1997, 82), which seeks to understand and causally explain processes with an important tempo-ral dimension, such as the formation and evolution of formal and informal social institutions. This trend has given rise to “historical institutionalism ” (for overviews see Hall and Taylor 1996; Immergut 1998; Robertso